P
US7398274B2ExpiredUtilityPatentIndex 81

Mention-synchronous entity tracking system and method for chaining mentions

Assignee: IBMPriority: Apr 27, 2004Filed: Apr 27, 2004Granted: Jul 8, 2008
Est. expiryApr 27, 2024(expired)· nominal 20-yr term from priority
Inventors:ITTYCHERIAH ABRAHAMJING HONGYANKAMBHATLA NANDAKISHORELUO XIAOQIANGROUKOS SALIM E
G10L 15/1822Y10S707/99943
81
PatentIndex Score
12
Cited by
11
References
6
Claims

Abstract

A Bell Tree data structure is provided to model the process of chaining the mentions, from one or more documents, into entities, tracking the entire process; where the data structure is used in an entity tracking process that produces multiple results ranked by a product of probability scores.

Claims

exact text as granted — not AI-modified
1. A program storage device embodying instructions executable by a processor to perform a method for entity tracking, the method comprising:
 inputting a document; 
 inputting an entity tracking query corresponding to the document; 
 determining a plurality of mentions and a plurality of objects in the document; 
 selecting each of the plurality of mentions in the document, one at a time, wherein a collection of mentions referring to an object of the plurality of objects forms an entity; 
 creating a partial entity for the selected mention and associating the partial entity with each existing partial entity, wherein each association is associated with a respective node on a tree graph; 
 linking the selected mention to each of the existing partial entities, wherein each linking is associated with a respective node on the tree graph; and 
 determining a score for each node of the tree graph, wherein each node corresponds to a different entity outcome; and 
 selecting a node of the tree graph as a response to the query according to the scores for the nodes, wherein a selected node is a response to the entity tracking query. 
 
   
   
     2. The program storage device of  claim 1  wherein the scores are cumulative along the chain of mentions in the tree graph. 
   
   
     3. The program storage device of  claim 1 , further comprising program nodes whose score falls below a threshold. 
   
   
     4. The program storage device of  claim 1 , further comprising selecting a subset from a plurality of nodes ranked by score. 
   
   
     5. The program storage device of  claim 1 , wherein a beam search algorithm is used to select a subset of nodes. 
   
   
     6. The program storage device of  claim 1  used for system training.

Cited by (0)

No later patents cite this yet.

References (0)

No backward citations on record.